> Continued > Fundamentally, on the ground, you're looking at economy of scale, both in terms of implementational scale and in terms of humanity collectively r

> Continued > Fundamentally, on the ground, you're looking at economy of scale, both in terms of implementational scale and in terms of humanity collectively responding to a problem and scale and slowly annealing (if I can borrow that from "simulated annealing") toward a solution that just happens to work. The macro level works out; the micro level looks terrible. Such is how this kind of thing works out; social scale/crowd dynamics are _really_ fascinating to watch at work. Node.JS happened, so suddenly JavaScript needs to run on servers and in browsers, and handle backend (Databases! Legacy stuff! FoxPro! How To Port Your Visual Basic Excel Macro To JavaScript! I just found two libraries for RS485!), along with frontend (Win32 event loop inspired Rube Goldberg machines!), and everything in between (......uhhhh...... all of GitHub?). The ecosystem and infrastructure - all the library code - has to be compatible with everything and everything else, across more domains than I even realized existed until I thought this through 5 minutes ago. The easiest way to work in the browser and also work on the backend apparently is to fragment everything out, because nobody's come up with a better way to stdlib JavaScript. The jQuery-esque approach of "this system works unbelievably well within its problem domain but doesn't generalize" effectively died when jQuery aged beyond its point of relevance - and because the plugin system was so locked-in to jQuery, all the plugins died with it. That type of approach is very very top-heavy and simply didn't scale. It made sense when all we had was browser [delivery] though, backend types hadn't gotten their mitts over everything and made compilation pipelines for everything, etc. Now that's happened, it makes more sense to `cat * > out.js` (well, in more LOC, anyway) because it just scales better. JS feels clunky, but once you look all the way down, I feel like the vertigo actually causes things to kinda make a tiny bit more sense and fit into place a bit better. Maybe. I find it not unreasonable to use the explanations above to justify the higher-level "why???" of the way things are. The duct tape and bubble gum isn't JS-specific. It's endemic to all of technology. Like with media, there are vested interests in things working out certain ways with security and whatnot. (My current favorite meme of this is http://web.archive.org/web/20160409181051/http://article.gmane.org/gmane.os.openbsd.misc/211963 http://web.archive.org/web/20160409181051/http://article.gma..., and I also found https://news.ycombinator.com/item?id=17160109 https://news.ycombinator.com/item?id=17160109 the other day and that was kind of very fun to read, and I also read the other day that the NSA discovered Heartbleed "very shortly" after it "appeared" in 2012 or whenever it was, which was a TIL of its own) And so it is that we use systems that operate at cross-purposes with how we think and incorporate high-friction impedance mismatches, and which raise the chances we'll make mistakes, which affect everyone: https://twitter.com/x0rz/status/865196993215442944 https://twitter.com/x0rz/status/865196993215442944, https://sites.google.com/site/testsitehacking/ https://sites.google.com/site/testsitehacking/ The above being said, I do completely understand your frustration with using Node. I have the same gripes with PHP, which I learned ages ago and Really Really need to move off of, but it just runs so incredibly well on older hardware and has zero warmup time... like 0.2ms on my 12 year old laptop... sigh. One day I'll find something, or maybe I'll make it myself. I'll leave you with three last links which I couldn't fit into the narrative, but which I think are relevant/related. Firstly this an interesting article that helped me properly understand Google: https://medium.com/p/e836451a959e https://medium.com/p/e836451a959e Next, I wanted to call out this NNTP post from 1995 (linked from "sent out via" in the above) showing storage investments: https://groups.google.com/forum/#!topic/mail.cypherpunks/4CDiW59hS88 https://groups.google.com/forum/#!topic/mail.cypherpunks/4CD... For context about cost in the above link: I found this very cute little document while randomly manual trawling (I love finding things not in Google's index) whose HTTP headers say it's from 1996: http://users.monash.edu.au/%7e%6a%6f%68%6e%6d/%77%65%62%73%65%72%76%65%72%32%2e%74%78%74 http://users.monash.edu.au/%7e%6a%6f%68%6e%6d/%77%65%62%73%6... (#1 I don't want its filename to be indexed just yet :P and #2 I want to do an experiment to see if the crawler resolves the escapes and finds the URL anyway) NB. If you have a better term than "new media", I'm curious. It's such an easy term to overload. The symbol picked for biohazard was designed from scratch to be meaningless so a meaning could be cultivated for/associated with it over time, and that cultivated meaning would be what "stuck". (I wonder how many other things this has happened with.) https://en.wikipedia.org/wiki/Hazard_symbol#Biohazard_symbol https://en.wikipedia.org/wiki/Hazard_symbol#Biohazard_symbol NB2. If you - say - accidentally hit F5 while writing a post... and you're on Linux, suddenly get really really interested in opening Chrome's task manager, identifying the renderer PID and SIGSTOPping it in as few seconds as you can. The first ~15% of the above text was recovered via gdb generate-core-file \o/ NB2a: (For completeness, if you accidentally close the tab, have working hibernation set up, and either go fishing in the swap partition on resume, or if you don't want to risk stepping on anything in swap on reboot, carefully boot something that won't automatically swapon.) NB3: $lists_i_am_on += 5000 :P